Stage Structure, Density Dependence and the Efficacy of Marine Reserves

The habitat requirements of fishes and other marine organisms often change with ontogeny, so many harvested species exhibit such extreme large-scale spatial segregation between life stages that all life stages cannot be protected within a single marine reserve. Nevertheless, most discussions of marine reserves have focused narrowly on single lifehistory events (e.g., reproduction or settlement) or a single life stage (e.g., adult or recruit). Instead, we hypothesize that an effective marine reserve system should often include a diversity of protected habitats, each appropriate to a different life stage. In such a case, the spatial configuration of habitats within reserves, and of separate reserves across larger spatial scales, may affect how marine resources respond to reserve design. We explored these issues by developing a mathematical model of a fish population consisting of two benthic life stages (juvenile and adult) that use spatially distinct habitats and examined the population’s response to various management scenarios. Specifically, we varied the sizes of reserves protecting the two life stages and the degree of coupling between juvenile and adult reserves (i.e., the fraction of the protected juvenile stock that migrates into the adult reserve upon maturation). We examined the effects when density dependence operated in only the juvenile stage, only the adult stage, or both. The results demonstrated that population stage structure and the nature of density dependence should be incorporated into the design of marine reserves but did not provide robust support for the general tenet that all life stages must be protected for an effective reserve system. The results indicated that biological considerations, alone, were insufficient for design of the optimal marine reserve. Instead, it was necessary to consider the value (e.g., socioeconomic or ecological) of each biological outcome; under some value functions, a mixed strategy (i.e., protecting both life stages) was best, whereas for others, the best solution focused on a single life stage. Resolving issues of marine reserve design, especially for stage-structured populations, will require more detailed study of stage-structured populations and a more explicit integration of biological and socioeconomic models. Marine reserves are often touted as a panacea for resource management. Presumed benefits include stock preservation and increased fisheries yields resulting from transport of larval production (and migration of older fishes) out of reserves (see, e.g., Plan Development Team, 1990; Roberts and Polunin, 1993; Clark, 1996; Lauck et al., 1998; but see also Allison et al., 1998; Bohnsack, 1998). The design of marine reserves is sometimes guided by model systems involving specific target species, such as gag grouper (Mycteroperca microlepis) (Koenig et al., this issue). This approach assumes that reserves will have similar benefits for other target species. Because their implementation can be difficult, both politically and economically, it is essential that reserves live up to expectations. They cannot simply be evaluated empirically after they are established. Indeed, tremendous resources can be wasted, and public confidence eroded, by a hit-and-miss approach. Mathematical models can provide initial insight into the implications of alternative strategies and can reveal how the implications depend on specific characteristics 676 BULLETIN OF MARINE SCIENCE, VOL. 66, NO. 3, 2000 of the managed system. Thus, theoretical investigations can help improve the design of marine reserves and clarify the type of empirical data needed to evaluate their efficacy. Most exploited fishes exhibit complex habitat requirements that change with ontogeny. The segregation of the life cycle of a ‘typical’ marine fish into a planktonic larval stage and a subsequent benthic stage is often recognized in existing models of marine reserves. Yet, because planktonic stages are not easily subject to management action, attention is generally focused on the benthic stage (see, e.g., Attwood and Bennett, 1995; Man et al., 1995). Within the benthic stage, however, many fishes continue to undergo ontogenetic changes in their habitat use and spatial distribution. Sometimes these changes are so dramatic that all benthic life stages could not possibly be included within any single reserve (see, e.g., Koenig et al., this issue). For example, in the Gulf of Mexico, gag typically recruit to estuarine seagrass beds, where they spend the first several months of their benthic life. Juvenile gag then migrate 10–30 mi offshore to hard-bottom reefs, where they spend an additional 2–6 yrs. Upon maturation, gag disperse further offshore (60–100 mi) to join other reproductive individuals at traditional spawning sites. Thus, gag use at least three potentially critical, and spatially segregated, benthic habitats at different stages in their life history (Hood and Schlieder, 1992; Ross and Moser, 1995; Coleman et al., 1996; Koenig and Coleman, 1998; Lindberg and Loftin, 1998; Koenig et al., this issue). Any of these spatially separated habitats might be considered “essential” sensu the Magnuson-Stevens Fishery Conservation and Management Act (Magnuson-Stevens Act, 1996) and worthy of a reserve. This sort of complex life cycle is not unique and probably occurs in many, if not most, harvested marine species (e.g., fishes, Koenig et al., this issue; crustaceans, Herrnkind et al., in press; the obvious exceptions are sessile invertebrates such as bivalves). As a result of these ontogenetic habitat shifts, populations become segregated into ecologically distinct life stages, giving rise to what we call ‘population stage structure’. Importantly, theoretical and empirical work in a variety of aquatic systems demonstrates that stage structure can have serious implications for population dynamics and the response of populations to environmental change (e.g., Crouse et al., 1987; Osenberg et al., 1992; Marschall and Crowder, 1996; Nisbet et al., 1996; Herrnkind et al., in press). Thus, the success of marine reserves is likely to depend critically on the life history of the target species, yet existing theory on marine reserves largely ignores the effect of habitat shifts and population stage structure (but see, e.g., Mangel, this issue; Stockhausen et al., this issue). This limitation is particularly noteworthy given the emphasis on habitat relations in the reauthorization of the Magnuson-Stevens Act (1996). Recognizing different stages in the benthic life phase refocuses the question of interest away from “should we establish a reserve?” toward “should we establish a reserve for this stage, or that stage, or should we set up multiple reserves each focused on a different life stage?” If multiple reserves are the best strategy, then we must ask whether it matters how these reserves are arranged in space (i.e., how distant they are from reserves protecting the other life stage). Again, we use gag to provide a hypothetical example. Movement from seagrass beds to offshore reefs may occur in such a way that particular seagrass beds primarily supply adults to particular reefs (e.g., those most directly offshore from the seagrass). This potential coupling between seagrass beds and reefs could be used in designing a marine reserve system. For example, seagrass and reef reserves could be set up such that seagrass reserves supply juveniles primarily to the reef reserves, or they could be set up such that the reef reserves receive juveniles that matured in nonreserve seagrass 677 ST. MARY ET AL. : STAGE-STRUCTURED MARINE RESERVES habitat. The benefits associated with reserve coupling will probably depend on the lifehistory features of the target species and the effects of density dependence. In this paper we present a stage-structured model in which we recognize two distinct benthic life stages, ‘juveniles’ and ‘adults’, and evaluate the efficacy of marine reserves designed to protect juveniles only, adults only, or both juveniles and adults. In the latter context, we examine the effect of coupling between the juvenile and adult reserves. We also consider how the results are affected by where in the life history density dependence operates most strongly, and we discuss further theoretical and empirical investigations motivated by our initial theoretical exploration.